Real-polarized genetic algorithm for the three-dimensional bin packing problem

Andre Homem Dornas, Flávio Vinícius Cruzeiro Martins, João Fernando Machry Sarubbi, Elizabeth Fialho Wanner

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

This article presents a non-deterministic approach to the Three-Dimensional Bin Packing Problem, using a genetic algorithm. To perform the packing, an algorithm was developed considering rotations, size constraints of objects and better utilization of previous free spaces (flexible width). Genetic operators have been implemented based on existing operators, but the highlight is the Real-Polarized crossover operator that produces new solutions with a certain disturbance near the best parent. The proposal presented here has been tested on instances already known in the literature and real instances. A visual comparison using boxplot was done and, in some situations, it was possible to say that the obtained results are statistically superior than the ones presented in the literature. In a given instance class, the presented Genetic Algorithm found solutions reaching up to 70% less bins.
Original languageEnglish
Title of host publicationGECCO '17: proceedings of the Genetic and Evolutionary Computation Conference
Place of PublicationNew York, NY (US)
PublisherACM
Pages785-792
Number of pages8
ISBN (Electronic)978-1-4503-4939-0
ISBN (Print)978-1-4503-4920-8
DOIs
Publication statusPublished - 15 Jul 2017
EventGenetic and Evolutionary Computation Conference, GECCO '17 - Berlin, Germany
Duration: 15 Jul 201719 Jul 2017

Conference

ConferenceGenetic and Evolutionary Computation Conference, GECCO '17
CountryGermany
CityBerlin
Period15/07/1719/07/17

    Fingerprint

Bibliographical note

-

Cite this

Dornas, A. H., Martins, F. V. C., Sarubbi, J. F. M., & Wanner, E. F. (2017). Real-polarized genetic algorithm for the three-dimensional bin packing problem. In GECCO '17: proceedings of the Genetic and Evolutionary Computation Conference (pp. 785-792). ACM. https://doi.org/10.1145/3071178.3071327